Intelligent Decision Support Systems 101 — The basics(Part 1)
I was recently working on a research project which required me to explore the use of decision support systems(DSS) and intelligent decision support systems(IDSS) in a context such as Agriculture. In this piece, I will be providing some general non-technical basics on Intelligent Support Systems(IDSSs).
During the past three decades, decision support systems have evolved from simple model-oriented systems to advanced multifunction entities. Decision support systems(DSS) are commonly defined as computerized programs or computer-based systems used to provide recommendations, projections, judgements, and course of action in an organization, system or a business(Prof. Vicky Sauter, 2002). Sprague and Watson(1996) defined a DSS as an interactive computer-based system that helps decision-makers use data and models to solve ill-structure, unstructured and semi-structured problems. DSSs can be classified in many different categories with the most common ones being Data-driven, model-driven, knowledge-driven, document-driven, communication driven, inter-and intra-organization DSS.
DSSs produce comprehensive reports by gathering and analyzing data. They provide the ability to conduct varying analysis with little programming effort and can be utilised by non-technical users. They are mostly used in business to drive faster and smarter decisions and projections based on objective data. They use contextual information, problem domain knowledge, decision-making experience, and historical data, to provide insights and proposed solutions to decision-makers based on the problem diagnosis.
The integration of Artificial Intelligence(AI) into traditional DSSs to create Intelligent Decision Support Systems(IDSSs) has been used to adequately model human problem-solving. Intelligent decision support systems (IDSSs), have been in use since the 1970s and are widely applied in various computer science applications for intelligent decision-making. Because real-world decision making requires the consideration and the analysis of multiple criteria features which are often conflicting in nature, decision-makers need more intelligent scientific methodologies to perform such complex evaluations.
As defined by Sarma in 1994, IDSSs are interactive computer-based systems which incorporate artificial intelligence techniques to provide solutions to semi-structured problems by using data, expert knowledge and models to support the decision-maker. They can also incorporate knowledge-based methodology to aid decision making by providing recommendations reflecting domain expertise. With the use of Machine Learning, IDSSs learn from previous experiences and improve with time, hence providing more accurate and sophisticated problem-oriented intelligent decision-making mechanisms that are continuously evolving. IDSSs also offer tremendous potential in support of well-defined tasks such as data conversion, information filtering, and data mining as well as supporting ill-structured tasks in dynamic cooperation. An IDSS is more of a cognitive rather than a technological system, given that even basic characteristics of intelligence cannot be captured in mechanistic(Malhotra et al., 2003).
IDSSs are essentially DSSs that make the extensive use of artificial intelligence techniques by effectively incorporating databases, model bases and intellectual resources to support decision making. They also tend to use expert systems technology to enhance the capabilities of decision-makers in understanding a decision problem and selecting an appropriate alternative. IDSSs offer full control to the user regarding information acquisition, evaluation and making the final decision. They operate under the assumption that the decision-maker is already familiar with the problem domain and the data required for the solution. It commonly presents a plethora of alternatives.
“IDSSs are essentially DSSs that make the extensive use of artificial intelligence techniques by effectively incorporating databases, model bases and intellectual resources to support decision making.”
The most important elements of intelligent decision making are exhaustive data, information and knowledge collection, extraction, analytics and rational decision making, and its adaptation to the changing environments. Machine learning algorithms and diverse programming paradigms are used to implement IDSSs by incorporating methodologies such as empirical evaluation and approaches such as multilabel learning, statistical and information-theoretic, landmarking, and complexity methods. This enables them to pick the most appropriate learning and prediction algorithms.
Compared to traditional methods, IDSSs are worthwhile because they have the ability to
- Learn from experience;
- Make sense of ambiguous and contradictory dataset;
- Provide appropriate and timely alternatives to a problem;
- Using reasoning to solve problems and inferring in rational ways;
- Deal with perplexing situations;
- Apply knowledge to understand or modify the environment;
- Recognise the relevance of various components of the decision-making process
Currently, IDSSs provide decision support via text analytics and mining-based DSSs, ambient intelligence and the internet of things-based DSSs; biometric-based DSSs; recommender, advisory and expert systems; GA-based DSS; fuzzy sets DSS; rough sets-based DSS; intelligent agent-assisted DSS; process mining integration to decision support, adaptive DSS; computer vision-based DSS; sensory DSS and robotic DSS(Arturas Kaklauskas, 2015). IDSS applications are developed in various areas such as in product development and planning; management decisions; enterprise and manufacturing industries.
